From: Trond Myklebust <trondmy@hammerspace.com>
To: "aglo@umich.edu" <aglo@umich.edu>
Cc: "linux-nfs@vger.kernel.org" <linux-nfs@vger.kernel.org>,
"Anna.Schumaker@netapp.com" <Anna.Schumaker@netapp.com>
Subject: Re: [PATCH 6/9] pNFS: Handle NFS4ERR_OLD_STATEID on layoutreturn by bumping the state seqid
Date: Thu, 12 Sep 2019 16:29:19 +0000 [thread overview]
Message-ID: <56f43f996e42e970e1633aec5a12c756bbb3e1c7.camel@hammerspace.com> (raw)
In-Reply-To: <CAN-5tyGOjTGft-rvqD7EzhCS9CK1dNKpUq9hRZTqs8ivLZ0_cg@mail.gmail.com>
On Thu, 2019-09-12 at 11:14 -0400, Olga Kornievskaia wrote:
> Hi Trond,
>
> Can you explain why are we just bumping the seqid by 1 instead of
> what
> it was currently using to update it to the current value which could
> be more than off by one? I'm just speculating that we'll see the same
> behavior that we'll get the ERR_OLD_STATEID incremented by one until
> the current value?
We only bump by 1 if we see that we're already above the seqid of the
current layout stateid. Otherwise, we sync to the value of that layout
stateid before retrying.
>
> On Tue, Sep 10, 2019 at 4:09 AM Trond Myklebust <trondmy@gmail.com>
> wrote:
> > If a LAYOUTRETURN receives a reply of NFS4ERR_OLD_STATEID then
> > assume we've
> > missed an update, and just bump the stateid.
> >
> > Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
> > ---
> > fs/nfs/nfs4proc.c | 2 +-
> > fs/nfs/pnfs.c | 18 ++++++++++++++----
> > fs/nfs/pnfs.h | 4 ++--
> > 3 files changed, 17 insertions(+), 7 deletions(-)
> >
> > diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
> > index a5deb00b5ad1..cbaf6b7ac128 100644
> > --- a/fs/nfs/nfs4proc.c
> > +++ b/fs/nfs/nfs4proc.c
> > @@ -9069,7 +9069,7 @@ static void nfs4_layoutreturn_done(struct
> > rpc_task *task, void *calldata)
> > server = NFS_SERVER(lrp->args.inode);
> > switch (task->tk_status) {
> > case -NFS4ERR_OLD_STATEID:
> > - if (nfs4_layoutreturn_refresh_stateid(&lrp-
> > >args.stateid,
> > + if (nfs4_layout_refresh_old_stateid(&lrp-
> > >args.stateid,
> > &lrp->args.range,
> > lrp->args.inode))
> > goto out_restart;
> > diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
> > index abc7188f1853..bb80034a7661 100644
> > --- a/fs/nfs/pnfs.c
> > +++ b/fs/nfs/pnfs.c
> > @@ -359,9 +359,10 @@ pnfs_clear_lseg_state(struct
> > pnfs_layout_segment *lseg,
> > }
> >
> > /*
> > - * Update the seqid of a layout stateid
> > + * Update the seqid of a layout stateid after receiving
> > + * NFS4ERR_OLD_STATEID
> > */
> > -bool nfs4_layoutreturn_refresh_stateid(nfs4_stateid *dst,
> > +bool nfs4_layout_refresh_old_stateid(nfs4_stateid *dst,
> > struct pnfs_layout_range *dst_range,
> > struct inode *inode)
> > {
> > @@ -377,7 +378,15 @@ bool
> > nfs4_layoutreturn_refresh_stateid(nfs4_stateid *dst,
> >
> > spin_lock(&inode->i_lock);
> > lo = NFS_I(inode)->layout;
> > - if (lo && nfs4_stateid_match_other(dst, &lo->plh_stateid))
> > {
> > + if (lo && pnfs_layout_is_valid(lo) &&
> > + nfs4_stateid_match_other(dst, &lo->plh_stateid)) {
> > + /* Is our call using the most recent seqid? If so,
> > bump it */
> > + if (!nfs4_stateid_is_newer(&lo->plh_stateid, dst))
> > {
> > + nfs4_stateid_seqid_inc(dst);
> > + ret = true;
> > + goto out;
> > + }
> > + /* Try to update the seqid to the most recent */
> > err = pnfs_mark_matching_lsegs_return(lo, &head,
> > &range, 0);
> > if (err != -EBUSY) {
> > dst->seqid = lo->plh_stateid.seqid;
> > @@ -385,6 +394,7 @@ bool
> > nfs4_layoutreturn_refresh_stateid(nfs4_stateid *dst,
> > ret = true;
> > }
> > }
> > +out:
> > spin_unlock(&inode->i_lock);
> > pnfs_free_lseg_list(&head);
> > return ret;
> > @@ -1475,7 +1485,7 @@ int pnfs_roc_done(struct rpc_task *task,
> > struct inode *inode,
> > *ret = -NFS4ERR_NOMATCHING_LAYOUT;
> > return 0;
> > case -NFS4ERR_OLD_STATEID:
> > - if (!nfs4_layoutreturn_refresh_stateid(&arg-
> > >stateid,
> > + if (!nfs4_layout_refresh_old_stateid(&arg->stateid,
> > &arg->range, inode))
> > break;
> > *ret = -NFS4ERR_NOMATCHING_LAYOUT;
> > diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h
> > index 3ef3756d437c..f8a38065c7e4 100644
> > --- a/fs/nfs/pnfs.h
> > +++ b/fs/nfs/pnfs.h
> > @@ -261,7 +261,7 @@ int pnfs_destroy_layouts_byfsid(struct
> > nfs_client *clp,
> > bool is_recall);
> > int pnfs_destroy_layouts_byclid(struct nfs_client *clp,
> > bool is_recall);
> > -bool nfs4_layoutreturn_refresh_stateid(nfs4_stateid *dst,
> > +bool nfs4_layout_refresh_old_stateid(nfs4_stateid *dst,
> > struct pnfs_layout_range *dst_range,
> > struct inode *inode);
> > void pnfs_put_layout_hdr(struct pnfs_layout_hdr *lo);
> > @@ -798,7 +798,7 @@ static inline void
> > nfs4_pnfs_v3_ds_connect_unload(void)
> > {
> > }
> >
> > -static inline bool nfs4_layoutreturn_refresh_stateid(nfs4_stateid
> > *dst,
> > +static inline bool nfs4_layout_refresh_old_stateid(nfs4_stateid
> > *dst,
> > struct pnfs_layout_range *dst_range,
> > struct inode *inode)
> > {
> > --
> > 2.21.0
> >
--
Trond Myklebust
Linux NFS client maintainer, Hammerspace
trond.myklebust@hammerspace.com
prev parent reply other threads:[~2019-09-12 16:29 UTC|newest]
Thread overview: 17+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-09-09 14:00 [PATCH 1/9] pNFS: Ensure we do clear the return-on-close layout stateid on fatal errors Trond Myklebust
2019-09-09 14:00 ` [PATCH 2/9] NFSv4: Clean up pNFS return-on-close error handling Trond Myklebust
2019-09-09 14:00 ` [PATCH 3/9] NFSv4: Handle NFS4ERR_DELAY correctly in return-on-close Trond Myklebust
2019-09-09 14:00 ` [PATCH 4/9] NFSv4: Handle RPC level errors in LAYOUTRETURN Trond Myklebust
2019-09-09 14:01 ` [PATCH 5/9] NFSv4: Add a helper to increment stateid seqids Trond Myklebust
2019-09-09 14:01 ` [PATCH 6/9] pNFS: Handle NFS4ERR_OLD_STATEID on layoutreturn by bumping the state seqid Trond Myklebust
2019-09-09 14:01 ` [PATCH 7/9] NFSv4: Fix OPEN_DOWNGRADE error handling Trond Myklebust
2019-09-09 14:01 ` [PATCH 8/9] NFSv4: Handle NFS4ERR_OLD_STATEID in CLOSE/OPEN_DOWNGRADE Trond Myklebust
2019-09-09 14:01 ` [PATCH 9/9] NFSv4: Handle NFS4ERR_OLD_STATEID in LOCKU Trond Myklebust
2019-09-16 19:37 ` Olga Kornievskaia
2019-09-11 20:13 ` [PATCH 8/9] NFSv4: Handle NFS4ERR_OLD_STATEID in CLOSE/OPEN_DOWNGRADE Olga Kornievskaia
2019-09-11 20:56 ` Trond Myklebust
2019-09-12 15:01 ` Olga Kornievskaia
2019-09-12 15:04 ` Olga Kornievskaia
2019-09-16 19:39 ` Olga Kornievskaia
2019-09-12 15:14 ` [PATCH 6/9] pNFS: Handle NFS4ERR_OLD_STATEID on layoutreturn by bumping the state seqid Olga Kornievskaia
2019-09-12 16:29 ` Trond Myklebust [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=56f43f996e42e970e1633aec5a12c756bbb3e1c7.camel@hammerspace.com \
--to=trondmy@hammerspace.com \
--cc=Anna.Schumaker@netapp.com \
--cc=aglo@umich.edu \
--cc=linux-nfs@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).